﻿* Encoding: UTF-8.

FREQUENCIES VARIABLES= irrelevant
  /ORDER=ANALYSIS.

USE ALL.
COMPUTE filter_$=(irrelevant = 0).
VARIABLE LABELS filter_$ 'irrelevant = 0 (FILTER)'.
VALUE LABELS filter_$ 0 'Not Selected' 1 'Selected'.
FORMATS filter_$ (f1.0).
FILTER BY filter_$.
EXECUTE.

FREQUENCIES VARIABLES= user_followers_count user_following_count user_status_count rt_count like_count
  /STATISTICS=STDDEV MINIMUM MAXIMUM MEAN  VARIANCE SKEWNESS SESKEW
  /FORMAT=NOTABLE
  /ORDER=ANALYSIS.

FREQUENCIES VARIABLES= query type date_created verified debunk overt_debunk debunk_expressed_suspicion auth 
  /ORDER=ANALYSIS.

* comparing the variable on the right

CROSSTABS
  /TABLES=date_created BY debunk
  /FORMAT=AVALUE TABLES
  /STATISTICS=CHISQ 
  /CELLS=COUNT COLUMN BPROP 
  /COUNT ROUND CELL.


CROSSTABS
  /TABLES=debunk BY verified
  /FORMAT=AVALUE TABLES
  /STATISTICS=CHISQ 
  /CELLS=COUNT COLUMN BPROP 
  /COUNT ROUND CELL.

* for both myths: debunk tweets used more authority appeals

CROSSTABS
  /TABLES=auth_type BY debunk
  /FORMAT=AVALUE TABLES
  /STATISTICS=CHISQ 
  /CELLS=COUNT COLUMN BPROP 
  /COUNT ROUND CELL.

CROSSTABS
  /TABLES=auth BY debunk
  /FORMAT=AVALUE TABLES
  /STATISTICS=CHISQ 
  /CELLS=COUNT COLUMN BPROP 
  /COUNT ROUND CELL.

* water_gargle hold_breath
* (like_count < 10000) & (rt_count < 10000) & 
* (debunk = 0) & (like_count < 10000) & (rt_count < 10000) & 

* hks table 2 filter = ((irrelevant = 0) &  (debunk = 0) & (like_count < 10000))

USE ALL.
COMPUTE filter_$=((irrelevant = 0) &  (debunk = 0) & (like_count < 10000)).
VARIABLE LABELS filter_$ 'irrelevant = 0 (FILTER)'.
VALUE LABELS filter_$ 0 'Not Selected' 1 'Selected'.
FORMATS filter_$ (f1.0).
FILTER BY filter_$.
EXECUTE.

* haven't done spec vs abs appeals for each query separately
* overall using an abstract auth appeal descreases chance of likes and rts, which is mostly driven by misinformation tweets
* Generalized Estimating Equations.
GENLIN rt_count BY auth_abs verified (ORDER=DESCENDING) with follower_count_convert
  /MODEL auth_abs verified follower_count_convert INTERCEPT=YES
 DISTRIBUTION=NEGBIN(1) LINK=LOG
  /CRITERIA METHOD=FISHER(1) SCALE=1 MAXITERATIONS=10000 MAXSTEPHALVING=5 PCONVERGE=1E-006(ABSOLUTE) 
    SINGULAR=1E-012 ANALYSISTYPE=3(WALD) CILEVEL=95 LIKELIHOOD=FULL
  /REPEATED SUBJECT=handle SORT=YES CORRTYPE=INDEPENDENT ADJUSTCORR=YES COVB=ROBUST 
    MAXITERATIONS=100 PCONVERGE=1e-006(ABSOLUTE) UPDATECORR=1
  /MISSING CLASSMISSING=EXCLUDE
  /PRINT CPS DESCRIPTIVES MODELINFO FIT SUMMARY SOLUTION (EXPONENTIATED).

